Currently different methods are used to select informative individuals for re-sequencing and genotype imputation. In this study we compared the utility of the recently described identification of key contributors (KCO) method with two commonly applied strategies, namely the identification of pedigree-based marginal gene contributions (PED) and the optimization of genetic relatedness (REL) and against animals selected at random (RAN). Based upon a simulated population structure (5,100 individuals and 10,000 SNPs) we show that, KCO provided the highest phasing (lowest switch error rates) and imputation accuracies (0.5% and 91.5%), followed by PED (2.6% and 88.1%), RAN (1.6% and 87.5%) and REL (5.4% and 87.0%) when including a maximum number of 100 individuals in the reference population. Furthermore, it was demonstrated that with the selection of key contributors especially the imputation accuracy (correlation between true and imputed genotype) of rare variants (minor allelic frequency <0.1) can be significantly increased by more than 10%. Therefore, we suggest to include the individual genetic contribution score in the decision criteria when selecting individuals for re-sequencing and genotype imputation.